Delta-sigma algorithmic analog-to-digital conversion

Delta-sigma modulation for analog-to-digital conversion resolves a number of bits logarithmic in the number of modulation cycles, and linear in modulation order. As an alternative to higher-order noise shaping, we present an algorithmic scheme that iteratively resamples the modulation residue, by feeding the integrator output back to the input. This yields a bit resolution linear in the number of cycles, similar to an algorithmic analog-to-digital converter. The scheme simplifies the design of the digital decimator to a single shifting counter, and avoids interstage gain errors in conventional algorithmic analog-to-digital converters. Experimental results from an integrated CMOS array of 128 converters show the utility of the design for large-scale parallel quantization in digital imaging and hybrid analog-digital computing.